strisunshinewentingwang/pytorch

Pull Requests Commits

Split test_c10d.py to test_c10d_common.py, test_c10d_gloo.py, test_c10d_nccl.py (#56598)

pbelevich committed 3 years ago

5cc75e46

Update Kineto submodule and use new metadata api (#56432)

Ilia Cherniavskii committed 3 years ago

d24314bd

[iOS GPU][Design] Support multiple tensors as outputs (#56072)

xta0 committed 3 years ago

1b872744

Revert D27866138: [ONNX] Redesign inplace conversion (#55033)

malfet committed 3 years ago

36828aa0

Disable Windows GPU testing (#56655)

malfet committed 3 years ago

a1299a28

Fix errors when initializing Linear with 0 in_features (#56505)

beningodfrey4 committed 3 years ago

df1dfd87

Reland of "D27708346: generate xla codegen in-tree" (#56601)

bdhirsh committed 3 years ago

76fbd755

Enable skipped test for c10::complex on CUDA >= 11.2 (#50227)

zasdfgbnm committed 3 years ago

0cc42809

[ONNX] Redesign inplace conversion (#55033) (#56173)

BowenBao committed 3 years ago

24ff92f7

Add standardOps match more input type in ORT (#53813) (#56172)

BowenBao committed 3 years ago

818ce1d0

make ProcessGroupDefaultTimeout the same as python (#56549)

wanchaol committed 3 years ago

43ad172c

make ProcessGroup.Options.timeout argument private in python (#56531)

wanchaol committed 3 years ago

a970e525

s/“pad”/"pad"/ in files introduced by #56065 (#56618)

malfet committed 3 years ago

6d7d36d2

Add new scheduled job to circle-ci workflow (#55182)

soulitzer committed 3 years ago

5dcc7ac3

Fixing error in jit cuda on ROCm: non-constant-expression cannot be n… (#55243)

amathews-amd committed 3 years ago

73eaa0a5

[static_runtime] fix num args for to_copy (#56441)

ajyu committed 3 years ago

e0be76fb

Add devices to TensorPipe options (#56405)

mrshenli committed 3 years ago

d83ae5d1

[7/n] [torch/elastic] Rename _Rendezvous to _RendezvousState (#56535)

cbalioglu committed 3 years ago

853112bb

[6/n] [torch/elastic] Reorder type definitions in dynamic_rendezvous.py (#56534)

cbalioglu committed 3 years ago

21d9bc24

[5/n] [torch/elastic] Introduce the delay utility function (#56533)

cbalioglu committed 3 years ago

df91eb92

[4/n] [torch/elastic] Fix the finalizer of PeriodicTimer (#56532)

cbalioglu committed 3 years ago

76ca1eee

[package] resolve `__import__` calls on export (#55153)

Lilyjjo committed 3 years ago

c244d1c5

add dtype information for input (#55358)

guotuofeng committed 3 years ago

28f52649

[PyTorch] Migrate add operators to borrow in TensorIteratorBase (#55691)

swolchok committed 3 years ago

6032ea03

[PyTorch] Support borrowing in/out Tensors in TensorIterator (#55690)

swolchok committed 3 years ago

01842d2b

[PyTorch] Always update op.current_dtype in TensorIteratorBase::set_output (#55940)

swolchok committed 3 years ago

7e8f078a

[PyTorch] Remove non-const TensorIterator::tensor() method (#55420)

swolchok committed 3 years ago

b79901f9

[PyTorch] Format generated structured kernels code better (#55258)

swolchok committed 3 years ago

26fc27cb

[PyTorch] Fix const correctness for resize native functions (#55351)

swolchok committed 3 years ago

1211bccc

Use absolute path for local linter (#56633)

davidriazati@fb.com committed 3 years ago

5e695b12

Newer Older